Unveiling

the
WISC-IV
Donna Rury Smith, Ed.D.
The Psychological
Corporation
Agenda
A broad look at the changes and why we made
them.
A closer look at the new and familiar subtests
Research design of the new WISC IV
Psychometric properties of the test
Interpretive considerations

Role of Intelligence Tests
The task of assessing a childs intelligence
necessarily involves more than simply obtaining
his or her scores. As Wechsler (1975) noted:
What we measure with tests is not what tests measure
not information, not spatial perception, not reasoning
ability. These are only a means to an end. What
intelligence tests measure is something much more
important; the capacity of an individual to understand
the world about him and his resourcefulness to cope with
its challenges.
Why Revise?
New research on cognitive abilities,
enormous literature on the WISC III
Demographic Shifts
e.g. Hispanic population changes (from
11% to 15%)
Regions (growth of the West/South at
expense of NE)
Flynn Effect
Revision Goals - I
Strengthen Four-Factor Model
Improve assessment of:
Fluid Reasoning
Working Memory
Processing Speed
Enhance clinical utility, and provide strong
evidence of clinical validity.

Revision Goals - II
Remove time-bonuses where possible.
Put speed where it belongs.
Improve Psychometric Properties
Subtest floors and ceilings
Remove potentially biased items
Link to measures of achievement, memory,
adaptive behavior, emotional intelligence,
giftedness and a cognitive process
instrument.


Changes

Updated theoretical foundations
Enhanced clinical utility
Increased developmental appropriateness
Improved psychometric properties
Increased user friendliness

Similarities
Vocabulary
Comprehension
Information
Word Reasoning
Block Design
Picture Concepts
Matrix Reasoning
Picture Completion
Coding
Symbol Search
Cancellation
Digit Span
Letter-Number
Sequencing
Arithmetic
VCI
WMI
PRI
PSI
FSIQ
Structure of WISC IV
Subtests
Core Subtests (10 of them)
Are administered when composite scores are
desired
Supplemental Subtests (5 of them)
Extend the range of cognitive skills sampled and
provide additional clinical information as they
enable the clinician to complete additional
discrepancy analysis. They can also be used as
substitutes for core subtests.

Composite Scores
Dual (or dueling?) IQ and Index structure gone.
Nowfour Indexes:
Verbal Comprehension = VCI
Perceptual Reasoning = PRI
Working Memory = WMI
Processing Speed = PSI

and FSIQ (consisting of 10 subtests)
Full Scale IQ
Stronger contributions of working memory and
processing speed
30% each VCI and PRI
20% each PS and WM
Composite Scores
Perceptual Reasoning Index (PRI)
Shift in emphasis from organization to reasoning
Emphasis on fluid reasoning in the perceptual domain
Subtests are highly g-loaded

Composite Scores
Working Memory Index
Essential component of fluid reasoning and
other higher order skills
Closely related to achievement and learning

See Fry & Hale, 1996; Perlow, Juttuso, & Moore, 1997;
Swanson, 1996

Composite Scores
Processing Speed Index
Dynamically related to mental capacity,
reading performance & development, and
reasoning by conservation of resources (e.g.,
efficiency)

See Fry & Hale, 1996; Kail, 2000; Kail & Hall,1994; Kail
& Salthouse, 1994; Berninger, 2001

Process Scores
In addition to the subtest and composite scores,
several additional process scores which provide
more detailed information about a childs
performance, are available.
No additional administration procedures are
requires to derive these scores.
Process scores can NEVER be substituted for
core or supplemental subtest scores in the
calculation of composite scores.
User Friendliness
Testing time reduced
Administration procedures simplified
Use of supplemental subtests for a core subtest
based on clinical need and appropriateness
Manual reorganization
Record Form reorganization

Content Changes
Deleted 3 subtests
Picture Arrangement
Object Assembly
Mazes
Retained, but revised, 10 subtests
Added 5 new subtests
New Subtests

Picture Concepts
Letter-Number Sequencing
Matrix Reasoning
Cancellation
Word Reasoning

Picture Concepts
For each item, the child is presented with 2 or 3
rows of pictures and chooses one picture from each
row to form a group with a common characteristic.
Measure of fluid reasoning and abstract
categorical reasoning (without verbal response).
Items progress from relatively concrete to more
abstract.
28 items



Letter-Number Sequencing
The child is read a sequence of numbers and
letters and recalls the numbers is ascending
order and the letters in alphabetical order.
Measure of working memory
Adapted from the WAIS-III (but new items)
Involves sequencing, mental manipulation,
attention, short-term auditory memory,
visuospatial imaging, and processing speed
Gets credit with either order
Matrix Reasoning
The child looks at an incomplete matrix and
selects the missing portion from 5 response
options. (35 items)
Measure of fluid reasoning and perceptual
organization
Reliable estimate of general intellectual ability
4 types of items to assess skills
Continuous and discrete pattern completion
Classification
Analogical reasoning
Serial reasoning


Cancellation
The child scans both a random and structured
arrangement of pictures and marks target
pictures within a specified time limit.
Measure of processing speed and visual selective
attention
2 forms (Random & Structured)
Forms share identical target locations
Targets are animals
Foils are common non-animal objects
Word Reasoning
The child is asked to identify the common concept
being described in a series of clues.
Measure of verbal comprehension, analogical
and general reasoning ability, verbal abstraction,
domain knowledge, the ability to integrate and
synthesize different types of information, and the
ability to generate alternative concepts.
Designed to measure fluid reasoning with
verbal material.
24 items




Updates
to Existing Subtests
Verbal Comprehension Subtests
Similarities
Vocabulary
Comprehension
Information
Word Reasoning
Similarities
23 items, 11 of which are new
Scoring criteria for all items were revised
Sample item has been revised to require a
credible response from the child before
beginning the subtest rather than the corrective
feedback provided on the WISC III.
The number of 1 point items was reduced from
5 to 2.
Age-defined start points were added.
In what way are a ____ and a ____ alike?
Vocabulary
Picture naming items in the stimulus book
provide more floor.
Vocabulary words are now displayed in
Stimulus Book, in addition to being read aloud.
36 items, including 4 new picture items and 32
verbal items (5 new, 27 retained)
Scoring criteria for all verbal items were revised.

Comprehension
21 items, 11 of them are new
Scoring criteria for all items were revised


Information
Is now a supplemental subtest
33 items, 11 of them are new
Scoring criteria for some retained items have
been modified


Perceptual Reasoning Subtests
Block Design
Picture Concepts
Matrix Reasoning
Picture Completion

Block Design
14 items, 11 were retained from WISC III, 3 were added
to extend the ceiling
Instructions have been shortened
Discontinue rule has been increased from 2 to 3
consecutive scores of 0
Time bonus scores restricted to the last 6 items (total
raw scores for many young children will not include any
time bonus points)
Process scores available for untimed performance
(BDN)


Picture Completion
Now a supplemental subtest
All new artwork
Items are designed to focus less on attention to
minor visual details and more on reasoning
Scoring criteria were revised and include
distinctions between those verbal responses that
require pointing and those that should be
awarded credit without pointing
38 items, including 13 new items
Working Memory Subtests
Digit Span
Letter-Number Sequencing
Arithmetic

Digit Span
Is now a core subtest and requires administration of both
Digits Forward and Digits Backward
8 items in Digit Span Forward (2 trials each) and 8 items
in Digit Span Backward (2 trials each)
15 trials were retained in Digit Span Forward, and 12
were retained in Digit Span Backward
a 2-digit sample item was added to DSB
process scores are available to compare DSF and DSB
and LDSF and LDSB
Arithmetic
Reduced math knowledge requirements while
increasing the working memory demands
34 items, 11 items involve the same math
calculation as items on WISC III, but verbatim
instructions have been revised.
Time bonuses are gone.
Task more like real-world working memory
demands


Processing Speed Subtests
Coding
Symbol Search
Cancellation

Coding
Verbatim instructions for both forms A and B
have been shortened to be more age
appropriate.
Additional process scores will be on the WISC
IV PI, but can test the limits now
Symbol Search
Search A and B were retained from WISC III,
but 15 items were added to Symbol Search B to
improve the ceiling for older children.
Verbatim instructions for both forms have been
shortened.
Dropped WISC-III Subtests
Mazes
Why?
Poor reliability
Little clinical utility
Infrequent use
Too much emphasis on Perceptual
Organization and motor skills (less consistent
with new model)


Dropped WISC-III Subtests
Picture Arrangement
Why?
Misinterpretation!
Poor Reliabilities
Less consistent with new theoretical
model

Dropped WISC-III Subtests
Object Assembly
Why?
Less consistent with new theoretical
model (more dependent on perceptual
organization than perceptual reasoning)
Dependence on motor skills
Dependence on speeded performance
Less examiner-friendly

Standardization
of the WISC-IV
WISC Standardization
Ages 6 16
Standardization Sample n = 2200
200 children at each age year
National sample stratified by sex, parent
education level, ethnicity and region.
Extensive validity studies with 15 clinical
groups.
Links to several major tests
See Tables 3.2-3.5 in Technical Manual for
percentage descriptions of Standardization
sample



Evidence of Reliability
Internal Consistency
Improved coefficients for retained subtests.
Significant improvement overall of subtests from
WISC III.
Average reliability coefficients (calculated with
Fishers z transformation) across age groups:
Verbal Comprehension .94
Perceptual Reasoning .92
Working Memory .92
Processing Speed .88 (based on split-half)
Full Scale .97
Special Group Reliabilities
Based on sample of 661 children in 16 clinical
groups. Table 4.2 in Tech Manual provides
coefficients by special group.
The majority of the subtest reliability coefficients
across special groups are similar or higher to
those coefficients reported for the normative
sample.
Average subtest reliability coefficients range
from a low of .82 on DSF to a high of .93 on LN
and on MR.
Standard Errors of Measurement
At subtest level overall average SEMs range
(across ages) from low of .97 on LN to a high of
1.38 on Cancellation.
At index level overall average SEMs range
(across ages) from low of 2.68 on Full Scale to a
high of 5.21 on Processing Speed.
As a result of relatively high reliability estimates
of the composite scores, the confidence intervals
tend to be small and can be found in Tables A.2-
A.6 in the Administration Manual.
Test-Retest Stability
Based on a sample of 243 children, with 18-27
from each of the 11 age groups.
Time interval ranged from 13 to 63 days, with a
mean interval of 32 days.
Table 4.4 in Tech Manual indicate that scores
possess adequate stability across time for all 5
age groups. The average corrected stability
coefficient is highest for Vocabulary (.92), good
(in the .80s) for BD, S, DS, CD, LN, MR, C, SS,
PC, I, WR; and adequate for all others (.70s).
Test-retest Gains
less pronounced on the Verbal Comprehension
and Working Memory subtests
score differences for combined age groups
(primarily due to practice effects)
VCI 2.1 points
PRI 5.2 points
WMI 2.6 points
PSI 7.1 points
FSIQ 5.6 points
Interscorer Agreement
Based on sample of 60 cases scored
independently by 4 scorers with no previous
WISC IV scoring experience
Interscorer reliabilities were
.98 for Similarities
.98 for Vocabulary
.95 for Comprehension
.96 for Information
.97 for Word Reasoning
Score Differences
A statistically significant difference between scores
refers to the likelihood that obtaining such a difference by
chance is very low if the true difference between the
scores is 0. The level of significance reflects the level of
confidence you can have that the difference s a true
difference.
The difference between scores required for significance
is computed from the standard error of measurement of
the difference.
Table B.1 in the Administration manual lists the
differences between index scores required for statistical
significance at the .15 and .05 levels of significance by
age groups and by overall standardization sample.
Frequency of Index Score
Differences
The prevalence or frequency of an observed score
difference in the general population is called the base
rate.
Sometimes the difference between an individuals index
scores is statistically significant but is not infrequent. The
statistical significance of differences between scores and
the rarity of the difference are two different issues and
have two different implications for test interpretation.
Base rates are reported in Table B.2 of the
Administration manual by overall standardization sample
and by ability level. It is also reported by the direction of
the difference.
Subtest Differences
The interpretation of the difference between a single
subtest score and the childs own mean score is an
intraindividual comparison. Strengths and weaknesses
identified in this way are relative to this childs own ability
level.
Table B.5 of the Administration manual reports the
minimum differences between a single scaled score and
the average scales scores of various groups of subtests
required for statistical significance at the .15 and .05
levels.
Remember a difference can be statistically significant
but not especially unusual. Table B.5 also provides data
on the estimated base rates of the general population.
Subtest Differences
A difference between scaled scores on a pair of subtests
(pairwise comparison) can be determined to be
statistically significant (use Table B.3) while base rates
for differences between scores are reported in Table B.4.
Intersubtest scatter reflects the variability of a childs
scaled scores across the subtests. Such variability is
often considered as diagnostically significant. Table B.6
provides cumulative percentages of intersubtest scatter
within various WISC IV composite scales.
Process score differences may be of particular interest
for clinical reasons. Table B.9 presents minimum scaled
score differences required for statistical significance.
Base rate data is reported in Table B.10.
Evidence of Validity
Relationship to Other Measures
Mean Scores for WISC IV and WISC III
N=244 children, aged 616, mean interval 28 days

WISC IV WISC III Corrected r
12

VCI/VIQ 103.0 105.4 .87
PRI/PIQ 103.9 107.3 .74
WMI/FDI 101.5 103.0 .72
PSI/PSI 102.7 108.2 .81
FSIQ/FSIQ 104.5 107.0 .89
VCI/VCI 102.9 106.0 .88
PRI/POI 103.9 106.9 .72
Relationship to Other Measures
Mean Scores for WISC IV and WPPSI III
N=182 children, aged 67, mean interval 22 days

WISC IV WPPSI III Corrected r
12

VCI/VIQ 100.5 100.2 .83
PRI/PIQ 102.6 102.0 .79
WMI 99.6
PSI/PSQ 103.1 104.4 .65
FSIQ/FSIQ 102.7 102.5 .89
GLC 97.1
Relationship to Other Measures
Mean Scores for WISC IV and WAIS III
N=198 children, aged 16, mean interval 22 days

WISC IV WAIS III Corrected r
12

VCI/VIQ 97.3 100.2 .86
PRI/PIQ 98.9 102.3 .76
WMI/WMI 98.7 97.7 .79
PSI/PSI 99.5 102.8 .77
FSIQ/FSIQ 98.5 101.6 .89
VCI/VCI 97.3 100.8 .85
PRI/POI 98.9 103.8 .73
Relationship to Other Measures
Mean Scores for WISC IV and WASI
N=260 children, aged 616, mean interval 29 days

WISC IV WASI Corrected r
12

VCI/VIQ 98.9 102.6 .85
PRI/PIQ 101.4 104.0 .78
WMI 100.7
PSI/PSI 98.4
FSIQ/FSIQ-4 100.2 103.6 .86
FSIQ/FSIQ-2 100.2 102.0 .83
Relationship to Other Measures
Mean Scores for WISC IV and WIAT II
N=550 children, aged 616, mean interval 12 days

VCI and Total Achievement .80
PRI and Total Achievement .71
WMI and Total Achievement .71
PSI and Total Achievement .58
FSIQ and Total Achievement .87

Relationship to Other Measures
See Table 5.15 in Technical manual
Picture Concepts correlations range from low (.30 for Oral
Expression) to moderate (.43 for Listening Comp)
Letter-Number Sequencing correlations were moderate
across WIAT II subtests ranging from .39 (Oral
Expression) to .60 (Math Reasoning)
Matrix Reasoning correlations ranged from .42 (Oral
Expression) to .59 (Math Reasoning)
Cancellation correlated minimally with all WIAT II subtests
Word Reasoning correlations ranged from .62 (Listening
Comprehension) to .42 (Oral Expression)
Relationship to Other Measures
Table 5.16 Correlations with Childrens Memory
Scales
Table 5.17 Correlations with Gifted Rating Scale
(School Form)
Table 5.18 Correlations with BarOn EQ
Table 5.195.20 Correlations with Adaptive
Behavior Assessment ScalesSecond Edition
Clinical Studies
Clinical Matched Controls
Each clinical case is matched demographically
to normal cases based on age, sex, ethnicity,
parent education level, and geographic region.
The mean of all normal cases that match each
clinical case is used.




Clinical Studies
Mental Retardation Mild WISC-IV WISC-IV PI
Mental Retardation Moderate

Learning Disabled Reading
Learning Disabled Reading, Writing
Learning Disabled Reading, Writing, Math
Learning Disabled Math
WISC-IV WISC-IV PI WIAT-II PAL

Receptive Language Disorder
Receptive/Expressive Language Disorder
WISC-IV WISC-IV PI CELF-4
Clinical Studies
ADD
ADD/LD Combined WISC-IV WISC-IV PI BROWN ADD

Motor Impaired WISC-IV
Hearing Impaired WISC-IV
Autism/Aspergers WISC-IV WISC-IV PI

TBI Open WISC-IV WISC-IV PI NEPSY D-KEFS
TBI Closed WISC-IV WISC-IV PI NEPSY D-KEFS

Gifted WISC-IV D-KEFS EQ GRS

Mental Retardation Study-1
Mild MR Matched Control
Mean SD Mean SD
VCI 67.1 9.1 98.7 12.5
PRI 65.5 10.3 98.7 15.2
WMI 66.8 11.1 99.4 13.8
PSI 73.0 11.6 98.3 13.5
FSIQ 60.5 9.2 99.2 13.6



Mental Retardation Study-2
Moderate MR Matched Control
Mean SD Mean SD
VCI 52.3 7.5 97.2 14.1
PRI 52.5 9.2 99.2 15.2
WMI 57.0 9.5 98.9 14.6
PSI 58.2 11.0 97.3 12.3
FSIQ 46.4 8.5 98.0 14.5

Closer look at MR
1. Must consider both intellectual functioning and adaptive behavior.
2. The prevalence of large and unusual discrepancies between
verbal and nonverbal composite scores has been shown to
decrease with decreasing levels of ability (Slate, 1995; Spruill,
1996, 1998; Wechsler, 1997). Thus, there appears to be less
variability in performance at both the composite (index) and the
subtest levels for children with MR than for children in general
population.
3. Children with MR also have relative strengths and weaknesses.
Study with WISC III suggested that children with Mild MR may
perform slightly better on measures of processing speed than on
measures of verbal or perceptual-organization abilities (Wechsler,
1991).

LD - Reading
RD Match Diff Signif Effect
VCI 91.9 100.9 9.0 p<.0001 .89
PRI 94.4 99.3 4.9 p<.0001 .48
WMI 87 99.8 12.8 p<.001 1.10
PSI 92.5 98.6 6.16 p<.001 .53
LD Reading & Writing
RWD Match Diff Signif Effect
VCI 94.8 101.3 6.5 p<.000 .49
PRI 98.0 101.0 3.1 p<.000 .25
WMI 90.2 100 9.8 p<.0001 .77
PSI 90.6 102 11.4 p<.000 .87
LD Math
MD Match Diff Signif Effect
VCI 93.2 99.3 6.1 p<.000 .61
PRI 87.7 97.2 9.6 p<.000 .80
WMI 92.9 99.7 6.7 p<.000 .56
PSI 90.6 95.6 5 p<.000 .36
Related Assessments
in Development
Co-Normed
with WISC-IV PI

WISC-IV PI provides a number of standardized
subtests and scores
Allows testing-of-limits supported by normative
data
Multiple-choice version of Verbal subtests and
Block Design
Spatial SpanSpatial Working Memory
Elithorn MazesExecutive Function
Frequency of error types












WISC-IV Spanish Edition
To be normed entirely in the U.S.
Items adapted from WISC-IV
Where adaptation is impossible,
parallel items have been created.
Standardization 2003
We NEED your help!
In Summary
WISC IV
The 4-factor model has been strengthened by
adding new subtests that measure fluid
reasoning, working memory and processing
speed.

The number of core subtests required (and,
therefore, testing time) has been reduced.

Floors, ceilings, and reliabilities improved on all
subtests.



WISC IV
Subtests and indexes with demonstrated clinical
utility are provided.
Usability of the test kit improved.
Easy to administer and carry!
Dedicated website.




Remember
Psychological assessment is a clinical
activity that employs test scores, but
only as one of the sources from which
an astute clinician develops a well-
integrated and comprehensive
psychological portrait of the child
examined.
Contact me at
Dr. Donna Rury Smith
2817 Berry Trace
Schertz, TX 78154
(210) 566-6812 office
1-800-228-0752 Et 4034 voice mail
donna_smith@harcourt.com